Haplotype phasing in single-cell DNA-sequencing data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell DNA sequencing necessitates whole-genome amplification (WGA) due to limited DNA input.
• WGA introduces significant biases, including uneven genome coverage and allele dropout, hindering genomic variant detection.

Purpose of the Study:

• To develop a method that utilizes WGA-induced biases for haplotype phasing in single-cell DNA sequencing.
• To improve the accuracy and length of haplotype phasing compared to existing read-based methods.

Main Methods:

• Developed a statistical test to detect concurrent allele dropout across single-nucleotide polymorphisms (SNPs) in multiple single cells.
• Implemented a haplotype assembly algorithm using allele dropout correlations to phase SNPs.
• Applied the algorithm to whole-genome and whole-exome sequencing data from neural cells.

Main Results:

• Achieved significantly longer haplotype blocks (median 10.2 kb for WGS, 9.2 kb for WES) compared to read-based phasing (312 bp and 41 bp, respectively).
• Demonstrated high accuracy with error rates below 2% for WGS and below 4% for WES.
• Showcased the algorithm's ability to predict SNP phasing with greater accuracy than read-alone phasing.

Conclusions:

• WGA amplification biases can be repurposed as a powerful signal for long-range haplotype phasing in single-cell genomics.
• The developed algorithm offers a substantial improvement in haplotype block length and accuracy, facilitating studies of rare alleles and allele-specific somatic aberrations.